Para-Hydrogen Induced Polarization without Incorporation of Para

Kenan Tokmic, Charles R. Markus, Lingyang Zhu, and Alison R. Fout . ... Simon B. Duckett, Gary G. R. Green, Joaquín López-Serrano and Adrian C. Whit...
2 downloads 0 Views 240KB Size
Download PDF
Inorg. Chem. 2009, 48, 663-670

Para-Hydrogen Induced Polarization without Incorporation of Para-Hydrogen into the Analyte Kevin D. Atkinson,† Michael J. Cowley,† Simon B. Duckett,*,† Paul I. P. Elliott,‡ Gary G. R. Green,§ Joaquı´n Lo´pez-Serrano,† Iman G. Khazal,† and Adrian C. Whitwood† Department of Chemistry, UniVersity of York, Heslington, York YO10 5DD, U.K., and York Neuroimaging Centre, The Biocentre, York Science Park, InnoVation Way, Heslington, York YO10 5DD, U.K. Received October 20, 2008

The cationic iridium complexes [Ir(COD)(PR3)2]BF4 (1a-c) (a, R ) Ph; b, R ) p-tolyl; c, R ) p-C6H4-OMe) react with parahydrogen in the presence of pyridine to give trans, cis, cis-[Ir(PR3)2(py)2(H)2]+ (2a-c) and small amounts of fac, cis-[Ir(PR3)(py)3(H)2]+ (3a-c), each of which exhibit polarized hydride resonances due to the magnetic inequivalence associated with the resultant AA“XX” spin system when 15N-labeled pyridine is employed. The pyridine ligands in 2 are labile, exchanging slowly into free pyridine with a rate constant of 0.4 s-1 for 2a at 335 K in a dissociative process where ∆Hq ) 134 ( 1 kJ mol-1 and ∆Sq ) 151 ( 5 J mol-1 K-1. Pyridine ligand exchange in 2 proves to be slower than that determined for 3. Parahydrogen induced polarization (PHIP) based on the hydride ligands of 2 and 3 is transferred efficiently to the 15N nuclei of the bound pyridine ligand by suitable insensitive-nuclei-enhanced-by-polarization-transfer (INEPT) based procedures. Related methods are then used to facilitate the sensitization of the free pyridine 15N signal by a factor of 120-fold through ligand exchange even though this substrate does not contain parahydrogen. This therefore corresponds to the successful polarization of an analyte by parahydrogen induced polarization methods without the need for the actual chemical incorporation of any parahydrogen derived nuclei into it.

Introduction Over the past two decades parahydrogen enhanced NMR spectroscopy1 has proved to be a powerful tool for the study of reaction mechanisms involving H2.2-4 During these studies several reaction intermediates have been detected and their kinetic fate monitored.5-10 Low concentration products * To whom correspondence should be addressed. E-mail: [email protected]. † University of York. ‡ Present Address: Department of Chemical & Biological Sciences, University of Huddersfield, Queensgate, Huddersfield HD1 3DH, U.K. § York Neuroimaging Centre. (1) Bowers, C. R.; Weitekamp, D. P. J. Am. Chem. Soc. 1987, 109, 5541– 5542. (2) Duckett, S. B.; Sleigh, C. J. Prog. Nucl. Magn. Reson. Spectrosc. 1999, 34, 71–93. (3) Natterer, J.; Bargon, J. Nucl. Magn. Reson. Spectrosc. 1997, 31, 293– 315. (4) Eisenberg, R.; Eisenschmid, T. C.; Chinn, M. S.; Kirss, R. U. A. Chem. Ser. 1992, 47–74. (5) Blazina, D.; Duckett, S. B.; Dyson, P. J.; Lohman, J. A. B. Angew. Chem., Int. Ed. Engl. 2001, 40, 3874–3876. (6) Blazina, D.; Duckett, S. B.; Dyson, P. J.; Lohmann, J. A. B. Chem.sEur. J. 2003, 9, 1046–1061.

10.1021/ic8020029 CCC: $40.75 Published on Web 12/15/2008

 2009 American Chemical Society

present in H2 addition equilibria have also been detected using this approach.11,12 This method, relies on the fact that parahydrogen (p-H2), the nuclear spin isomer of dihydrogen with the spin configuration Rβ-βR, can be easily isolated. Reactions with p-H2 then lead to products that are initially formed with a nuclear spin distribution that at best reflects that of the original dihydrogen molecule. In this way nonBoltzmann spin distributions can be produced in such reaction products which lead to significant signal enhancements in their 1H NMR spectra. It is for this reason that (7) Colebrooke, S. A.; Duckett, S. B.; Lohman, J. A. B.; Eisenberg, R. Chem.sEur. J. 2004, 10, 2459–2474. (8) Duckett, S. B.; Newell, C. L.; Eisenberg, R. J. Am. Chem. Soc. 1994, 116, 10548–10556. (9) Lo´pez-Serrano, J.; Duckett, S. B.; Lledo´s, A. J. Am. Chem. Soc. 2006, 128, 9596–9597. (10) Lopez-Serrano, J.; Duckett, S. B.; Aiken, S.; Lenero, K. Q. A.; Drent, E.; Dunne, J. P.; Konya, D.; Whitwood, A. C. J. Am. Chem. Soc. 2007, 129, 6513–6527. (11) Duckett, S. B.; Eisenberg, R. J. Am. Chem. Soc. 1993, 115, 5292– 5293. (12) Duckett, S. B.; Mawby, R. J.; Partridge, M. G. Chem. Commun. 1996, 383–384.

Inorganic Chemistry, Vol. 48, No. 2, 2009

663

Atkinson et al. parahydrogen induced polarization (PHIP) has facilitated the characterization of many species that would otherwise be unobservable by NMR spectroscopy. The best example of this enhancement is provided by the hydride resonances of the complex [Ru(dpae)(CO)2(H)2] (dpae ) bis(diphenylarsino)ethane) when it is generated by photolysis of [Ru(dpae)(CO)3] under a p-H2 atmosphere.13,14 The hydride resonances were shown to yield the maximum theoretical signal enhancement of 31,200 at 9.4 T. This situation actually corresponds to the generation of a pure singlet spin state and therefore means that every complex molecule contributes to the detected NMR signal. New insights into reactivity that have resulted from this effect include a more detailed delineation of the role of ruthenium clusters and their fragmentation products in hydrogenation catalysis.15 Other key species have been detected in cobalt and iridium catalyzed hydroformylation processes through the incorporation of the p-H2 label into alkyl and acyl ligands within the ligand sphere.16-18 Similarly, several palladium alkyl and vinyl complexes that play a direct role in alkyne hydrogenation have been detected, characterized, and studied.10,19 Many of these results have been reviewed.20,21 The detection of transition metal dihydride complexes through their hydride resonances is often facilitated because they appear in a normally vacant part of the 1H NMR spectrum, typically between -5 and -25 ppm. The polarization afforded by the PHIP effect therefore allows suitable metal dihydride complexes to act as sensors for substrates in low concentration through the detection of such signals. Here, the chemical shift of the hydride signal is diagnostic of the substrate. We recently reported studies dealing with the detection and dynamics of pyridine coordination derivatives of Wilkinson’s complex and its tribenzylphosphine and tricyclohexylphosphine analogues as part of such a study.22 More significantly, when it was extended to the receptor complex [Ir(H)2Cl(PPh3)2] the binding of purine and adenine substrates demonstrated that their detection at pico-mole levels was possible through such a hydride reporter.23 In these and other studies, the hydride-based polarizations have been used to illuminate the coordination sphere of metal (13) Anwar, M. S.; Blazina, D.; Carterer, H.; Duckett, S. B.; Halstead, T. K.; Jones, J. A.; Kozak, C. M.; Taylor, R. J. K. Phys. ReV. Lett. 2004, 93, 040501–040504. (14) Blazina, D.; Duckett, S. B.; Halstead, T. K.; Kozak, C. M.; Taylor, R. J. K.; Anwar, M. S.; Jones, J. A.; Carteret, H. A. Magn. Reson. Chem. 2005, 43, 200–208. (15) Blazina, D.; Duckett, S. B.; Dyson, P. J.; Lohman, J. A. B. Dalton Trans. 2004, 2108–2114. (16) Godard, C.; Duckett, S. B.; Henry, C.; Polas, S.; Toose, R.; Whitwood, A. C. Chem. Commun. 2004, 1826–1827. (17) Godard, C.; Duckett, S. B.; Polas, S.; Tooze, R. P.; Whitwood, A. C. J. Am. Chem. Soc. 2005, 127, 4994–4995. (18) Permin, A. B.; Eisenberg, R. J. Am. Chem. Soc. 2002, 124, 12406– 12407. (19) Lopez-Serrano, J.; Lledos, A.; Duckett, S. B. Organometallics. 2008, 27, 43–52. (20) Blazina, D.; Duckett, S. B.; Dunne, J. P.; Godard, C. Dalton Trans. 2004, 2601–2609. (21) Duckett, S. B.; Blazina, D. Eur. J. Inorg. Chem. 2003, 16, 2901– 2912. (22) Zhou, R. R.; Aguilar, J. A.; Charlton, A.; Duckett, S. B.; Elliott, P. I. P.; Kandiah, R. Dalton Trans. 2005, 3773–3779. (23) Wood, N. J.; Brannigan, J. A.; Duckett, S. B.; Heath, S. L.; Wagstafft, J. J. Am. Chem. Soc. 2007, 129, 11012–11013.

664 Inorganic Chemistry, Vol. 48, No. 2, 2009

complexes by polarization transfer. An early example of such a study dealt with the addition of p-H2 to Rh(PMe3)4Cl and Rh(PMe3)3Cl, where the dihydride products were used to demonstrate that two-dimensional NMR techniques such as COSY, HSQC, HMQC, and NOESY can be used with p-H2.24 Other one-dimensional methods of characterizing hydrogenation products have been developed by Bargon, for example, using a number of modified insensitive-nuclei-enhanced-by-polarization-transfer (INEPT) experiments to obtain 13C spectra of 1,4diphenylbut-1-en-3-yne, formed via the hydrogenation of 1,4-diphenylbutadiyne.25 Here we report on p-H2 studies dealing with the interaction of cationic dihydride iridium phosphine complexes and pyridine. The complexes employed here are based on the well-known [Ir(COD)L2]BF4 (COD ) 1,3- cyclooctadiene and L ) phosphine) system of Crabtree.26 Groups have prepared and conducted very extensive studies on the reactivity of these and the related complexes [IrH2S2L2]BF4 and [IrH2(ol)2L2]BF4 (S ) solvent; ol ) olefin) in both hydrogenation and dehydrogenation reactions. Related [IrH2S3L]BF4 complexes have also proved to be active in a variety of situations including alkyne dimerization and hydrogenation.27,28 Our experiments demonstrate that polarization transfer from the hydride ligands to the 15 N center of pyridine in [IrH2S2L2]BF4 (S ) pyridine) is possible. We also demonstrate that the PHIP effect can be used to enhance the signal strength of the 15N resonance of free pyridine via ligand exchange. This contrasts with the situation found in other, well-reported approaches, to polarization transfer with p-H2, where the NMR signature of organic materials is enhanced through their formation in a hydrogenation reaction involving an unsaturated version of the substrate.2-4 This report therefore presents a significant new development in the in situ preparation of hyperpolarized materials. Results and Discussion Reactivity of [Ir(COD)(PPh3)2]+ Toward H2 and Pyridine. When H2 was bubbled through a solution of [Ir(COD)(PPh3)2]BF4 (1a) in dichloromethane, in the presence of a 2-fold excess pyridine, complex [Ir(PPh3)2(py)2(H)2]+ (2a) was generated. Complex 2a was isolated, and appropriate NMR based spectroscopic data obtained to confirm its identity on the basis of agreement with previously described data of Rosales et al.29 The NMR parameters for 2a can be found in Table 1. 2a was then generated from 1a in CD3OD rather than CD2Cl2 by adding an 8-fold excess of pyridine and H2 to this solution at 330 K. When the time course of this reaction was monitored by in situ 1H NMR spectroscopy, the (24) Messerle, B. A.; Sleigh, C. J.; Partridge, M. G.; Duckett, S. B. Dalton Trans. 1999, 1429–1435. (25) Haake, M.; Natterer, J.; Bargon, J. J. Am. Chem. Soc. 1996, 118, 8688– 8691. (26) Crabtree, R. H.; Mellea, M. F.; Mihelcic, J. M.; Quirk, J. M. J. Am. Chem. Soc. 1982, 104, 107–113. (27) Sola, E.; Navarro, J.; Lopez, J. A.; Lahoz, F. J.; Oro, L. A.; Werner, H. Organometallics. 1999, 18, 3534–3546. (28) Navarro, J.; Sagi, M.; Sola, E.; Lahoz, F. J.; Dobrinovitch, I. T.; Katho, A.; Joo, F.; Oro, L. A. AdV. Synth. Catal. 2003, 345, 280–288. (29) Rosales, M.; Gonzalez, T.; Atencio, R.; Sanchez-Delgado, R. A. Dalton Trans. 2004, 2952–2956.

Para-Hydrogen Induced Polarization Table 1. NMR Data for Complexes 2a-c and 3a-ba 1

complex

H/δ

31

P/δ

13

C/δ

N/δb

15

-21.64 (2JPH )17.4 Hz, 2JNH ) 20, JHH ) -6.4) PPh3 7.37 (m, JHH ) 8.3 Hz, ortho), 7.26 δ 23.0 (s) 131.7 (1JPC 16 Hz, ipso), 133.7 (o), (t, J ) 8.3, meta), 7.36 (m, para) 128.0 (m), 130.0 (p) Py 8.05 (d, J ) 7.0 Hz, o), 6.78 (t, J ) 155.8 (o), 125.3 (m), 136.1 (p) δ -62.2 7.0, m), 7.49 (t, J ) 7.7, p) 2b H -21.47 (t, 2JPH ) 17.6 Hz, 2JNH ) 24, 2JHH ) -6.5) P(p-Tol)3 7.18 (dd, 7.7 and 6.8 Hz, o), 7.06 δ 22.8 (s) 133.5 (o), 128.6 (m), 140.5 (p), (d, 7.7, m), 2.33 (s, CH3) 19.9 (s, CH3) Py 8.02 (d, J ) 5.6 Hz, o), 6.77 (t, J ) 155.9 (o), 125.3 (m), 131.7 (p) δ -62.8 7.0, m), 7.50 (t, J ) 6.5, p) 2c H -21.73 (2JPH ) 17.1 Hz, 2JNH ) 23.3, 2JHH ) -6.5) P(p-C6H4OMe)3 7.22 (dt, J ) 8.0, 5.4 Hz, o), 6.79 18.2 (s) 127.0 (o), 105.7 (m), 153.4 (p), (d, J ) 8.0 Hz, m), 3.77 (s, OCH3). 115.55 (i), 46.58 (OCH3) Py 8.05 (d, J ) 6 Hz, o), 6.82 (t, J ) 148.07 (s, o), 117.6 (s, m), 128.7 (s, δ -62.0 7.0, m), 7.54 (t, J ) 7.0, p) p) 2 2 3a H -21.8 ( JPH ) 22.7 Hz, JNH ) 22, JHH ) -6.6) PPh3 7.40 (m, JHH ) 8.3 Hz, ortho), 7.23 12.6 (d, 2JPN ) 42 Hz) 133.1 (o), 127.8 (m), 129.8 (p) (t, J ) 8.3, meta), 7.33 (m, para) trans-Py 8.48 (d, J ) 7.0 Hz, o), 7.15 (t, 7.0, 154.5 (s, o), 125.8 (s, m), 136.7 (s, -48.2 m), 7.73 (t, 7.7, p) p) cis-Py 8.90 (d, J ) 7.0 Hz, o), 7.40 (t, J ) 155.6 (s, o), 124.1 (s, m), 137.0 (s, -64.0 7.0, m), 7.96 (t, J ) 7.7, p) p) 3b H -21.75 (2JPH ) 24 Hz, 2JNH ) 22) P(p-Tol)3 7.21 (t, J ) 8.3 Hz, o), 7.02 (t, J ) 9.74 (d, 2JPN ) 43 Hz) 133.15 (o), 128.14 (m), 19.78 (CH3) 8.3,m), 2.32 (s, CH3) trans-Py 8.43 (d, J ) 7.0 Hz, o), 7.13 (t, 7.0, 154.65 (o), 125.58 (m), 136.8 (p) -48.7 m), 7.78 (t, 7.7, p) cis-Py 8.97 (d, J ) 7.0 Hz, o), 7.39 (t, J ) 154.9 (o), 125.83 (m), 137.9 (p) -64.6 (d, 2JNP ) 43 Hz) 7.0, m), 7.97 (t, J ) 7.7, p) 3c H -21.71 (2JPH ) 23.2 Hz, 2JNH ) 22) P(p-C6H4OMe)3 7.28 (dd, J ) 8.9. 10.9 Hz, o), 6.78 6.8 (d, 2JPN ) 49 Hz) 134.6 (o), 113.1 (m), 160.9 (CO), (dd, J ) 1.7, 9.0 Hz, m), 3.79 (s, 123.4 (d, JCP ) 66 Hz) 54.4 (CH3) OCH3) trans-Py 8.47 (d, J ) 4.9 Hz, o), 7.18 (m, 154.7 (o), 125.65 (m), 136.84 (p) -48.3 6.6, m), 7.80 (tt, 7.7, 1.6, p) cis-Py 8.88 (m, J ) 6.0 Hz, o), 7.31 (t, J 155.4 (o), 125.9 (m), 137.8 (p) -64.1 (d, 2JNP ) 49 Hz) ) 7.1, m), 7.94 (tt, J ) 7.8,1.6, p) a All chemical shifts were measured at 295 K in CD3OD solution. b Relative to 15N-pyridine ) 0 ppm. trans-Py ) trans to H, cis-Py ) cis to H. 2a

H

Scheme 1. Reaction of 1 with H2 in the Presence of Excess Pyridine for Form Complexes 2 and 3 (R) Ph (a), p-tolyl (b) and p-C6H4-OMe (c))

formation of 2a was indicated by the observation of a dominant hydride signal at δ -21.64 (t, JHP ) 17.4 Hz) after a few minutes. The structure of 2a is indicated in Scheme 1, and it should be noted that the two hydride ligands are chemically equivalent and lie trans to pyridine. The formation of a small amount of a second complex was also indicated in this NMR spectrum, as evidenced by a second hydride signal at δ -21.8. This second hydride signal, due to species 3a, appears as a simple phosphoruscoupled doublet of 22.7 Hz. The presence of a single phosphorus-hydride splitting suggests that 3a is a monophosphine complex, and the small value of the JPH coupling indicates that the phosphine ligand is arranged cis to the hydride groups. The 31P resonance of the phosphorus center in 3a was located in an 1H-31P HMQC experiment as a

singlet at δ 12.6. When 15N-labeled pyridine was employed in this reaction the corresponding 31P resonance now appears as a nitrogen coupled doublet of 42 Hz. This is indicative of a trans phosphine-pyridine ligand arrangement in this monophosphine complex. More importantly, when this reaction was repeated with 15 N-labeled pyridine and p-H2 rather than H2 the hydride resonances of both 2a and 3a appear PHIP enhanced. These signals are illustrated in Figure 1 where the upper trace exhibits both 31P and 15N splittings while the lower trace is 31 P decoupled thereby revealing apparent 15N splittings of 20 and 22 Hz, respectively, in addition to the antiphase components, which are separated by -6.4, and -6.6 Hz, respectively. This observation is expected for the hydride resonance of the 15N-labeled isomer of the dihydride complex 2a since it now belongs to a second order spin system of the type AA“XX”. It is the 15N label that makes the hydride ligands magnetically inequivalent and therefore PHIP active. Indeed, when a 1H NMR spectrum is recorded that is 15N decoupled the second order effect is quenched and the PHIP enhancement disappears. Inorganic Chemistry, Vol. 48, No. 2, 2009

665

Atkinson et al.

Figure 2. Molecular structure of the cation 2a (ORTEP). Selected bond distances [Å] and angles (deg): Ir(1)-N(2) 2.177(2), Ir(1)-N(1) 2.196(2), Ir(1)-P(1) 2.2987(6), Ir(1)-P(2) 2.3038(6), N(2)-Ir(1)-N(1) 92.94(8), N(2)-Ir(1)-P(1) 92.89(6), N(1)-Ir(1)-P(1) 100.45(5), N(2)-Ir(1)-P(2) 97.82(6), N(1)-Ir(1)-P(2) 91.61(5), P(1)-Ir(1)-P(2) 163.44(2). Figure 1. (top) 1H NMR spectrum of the hydride region of a sample containing [Ir(COD)(PPh3)2]+ in CD3OD with 15N-pyridine under p-H2, (bottom) 1H{31P} NMR spectrum.

The hydride signals for 3a show a similar enhancement profile to those of 2a when the 1H NMR spectrum is recorded with 31P decoupling (Figure 1 (bottom)) and also disappear on 15 N decoupling. These observations are therefore indicative of similar ligand arrangements in both 2a and 3a. 3a is therefore confirmed as the dihydride monophosphine tris-pyridine complex fac, cis-[Ir(PPh3)(py)3(H)2]+. It is formed by exchange of one of the phosphine ligands in 2a with pyridine. The related complex fac, cis-[Ir(NCMe)3(PiPr3)(H)2]+ has been described in the literature.27 One interesting feature that manifests itself in Figure 1 is the relative size of the PHIP enhanced hydride resonances of 3a and 2a. In the PHIP enhanced spectra, the hydride signal for 3a is much larger than that seen for 2a even though 2a has the higher real concentration. This effect is reflected in the corresponding 31P{1H} NMR spectrum where no signal for 3a is visible after 128 scans, although that for 2a is. This situation also agrees with observations when a 9-fold excess of unlabeled pyridine is present since the relative hydride resonance intensities now prove to be 1 to 0.04, respectively. This information suggests therefore that the hydride ligands of complex 3a exchange with p-H2 at a much greater rate than those of 2a. The presence of pyridine within the coordination spheres of complexes 2a and 3a was also confirmed by 15N-based NMR spectroscopy. 1H-15N HMQC spectra of the 15Nlabeled samples contain correlations between the hydride signals of 2a and 3a and resonances in the 15N domain at δ -62.2 and -64.0, respectively. These 15N signals arise from the nitrogen center of the bound pyridine ligands that lie trans to hydride. The NMR parameters for these two complexes are listed in Table 1. The following text describes more fully how complexes 2a and 3a were characterized by NMR spectroscopy. First, when a 1D NOE experiment was recorded, the hydride signal at δ -21.64 for 2a proved to connect with proton signals at δ 8.05 and 7.37, where the latter signal couples strongly to 31P and therefore arises from an ortho phenyl proton and the former signal belongs to the ortho protons of bound pyridine. When the ortho pyridine protons at δ 8.05 are interrogated in a similar way in a 2D experiment, NOE connections to the corresponding meta protons at δ 6.78 and the ortho phenyl protons of the phosphine at δ 7.37 are visible. The meta pyridine protons

666 Inorganic Chemistry, Vol. 48, No. 2, 2009

proved to connect to the corresponding para protons which resonate at δ 7.49 in the same experiment. In contrast, the δ 7.37 signal for the phenyl ring connected to a signal at δ 7.26 due to a meta proton which connected in turn to a para proton signal which appears at δ 7.36. The remaining heteronuclear signals of 2a were located by HMQC methods, and the corresponding signals for 3a were attributed by a similar procedure. When 2 µL of 15N-labeled pyridine was added to a sample in which 2a and 3a had been generated previously using pyridine where the 15N label was present at isotopic abundance, only small levels of PHIP enhancement are observed for the corresponding hydride resonance of 3a and essentially no enhancement is observed for the corresponding hydride resonance of 2a. This suggests that the pyridine ligands of both these complexes do not exchange rapidly with free pyridine at 295 K. The enhancement observed for 3a is, however, compatible with it having a higher rate of exchange than that of 2a. X-ray Structure of [Ir(PPh3)2(py)2(H)2]+ (2a). The crystal structure of 2a with a BF4- counterion is illustrated in Figure 2. Previously Rosales et al. reported the same structure with the PF6- counterion.29 We note that with BF4- the P-Ir-P bond angle is 163.44(7)° but with PF6- it is 161.00(4)°. Furthermore, the reported Ir-N distances are 2.212(6) and 2.193(7) while here they are 2.177(2) and 2.196(2). While there are, however, no significant deviations between the Ir-P bond lengths, the N-Ir-N bond angles also vary from 92.94(8)° to 96.3(3)°, respectively. These changes may simply reflect crystal packing effects, but it should be noted that the ability of the precursor complexes as hydrogenation catalysts depends critically on the identity of the anion.30-32 Reactions of [Ir(COD)(P{p-tolyl}3)2]+ (1b) and [Ir(COD)(P{C6H4-4-OMe}3)2]+ (1c) with Pyridine and H2. To probe the effect of the identity of the phosphine on this reaction, analogues of 2a and 3a containing P{p-tolyl}3 and P{C6H4-pOMe}3 were generated, in situ, using the directly related precursors [Ir(COD)(P{p-tolyl}3)2]+, 1b, and [Ir(COD)(P{C6H4(30) Crabtree, R. H.; Anton, D. R.; Davis, M. W. Ann. N.Y. Acad. Sci. 1983, 415, 268–270. (31) Crabtree, R. H.; Demou, P. C.; Eden, D.; Mihelcic, J. M.; Parnell, C. A.; Quirk, J. M.; Morris, G. E. J. Am. Chem. Soc. 1982, 104, 6994– 7001. (32) Chodosh, D. F.; Crabtree, R. H.; Felkin, H.; Morris, G. E. J. Organomet. Chem. 1978, 161, C67–C70.

Para-Hydrogen Induced Polarization

p-OMe}3)2]+, 1c, respectively. The NMR data for these species was then collected and assigned in a similar way to that described for 2a. This information is presented in Table 1. Reactivity of [Ir(PPh3)2(py)2(H)2]+ (2a), [Ir(P{p-tolyl}3)2(py)2 (H)2]+ (2b) and [Ir(P{C6H4-4-OMe}3)2(py)2(H)2]+ (2c). Hydride Ligand Exchange with Dihydrogen. To compare the relative ability of 2 to undergo H2 exchange, two samples, one containing 1a and 1b and the second 1a and 1c, were prepared in CD3OD. They were then exposed to a 15-fold excess of 15Npyridine and p-H2 and then examined by 1H NMR spectroscopy at 295 K. Neither sample showed any PHIP enhancement in the corresponding hydride resonances of 2 or 3 although they were detectable. The samples were then warmed to 330 K and re-examined. At this point, the hydride signals of 2 and 3 were polarized and a series of 1H, 1H{31P}, integrateable 31P{1H} and 2D HMQC data sets were recorded. These spectra demonstrated that both samples contained signals for the corresponding mixed phosphine dihydride complex. These were [Ir(PPh3)(P{p-tolyl}3)(py)2(H)2]+ and [Ir(PPh3)(P{p-C6H4OMe}3)(py)2(H)2]+ respectively. NMR data for these complexes can be found in the Experimental Section. Key points of interest in these data include the observation of two distinct phosphorus signals for these species which exhibit a strong and mutual PP coupling of around 350 Hz. Furthermore, the dominant hydride signals detected in both samples under PHIP conditions actually corresponded to those of the mono phosphine complexes 3a, 3b, and 3c. The relative PHIP enhanced hydride signal intensities for 2 obtained from these samples were then determined, and then normalized by dividing them by the relative 31P integral values for these species. The normalized hydride signals intensity ratios determined in this way proved to be 1:1.42: 1.9:2.32:3.60 for the species 2a/2b/2c/[Ir (PPh3)(P{ptolyl}3)(py)2(H)2]+/[Ir(PPh3)(P{C6H4-4-OMe}3)(py)2(H)2]+, respectively. It should be noted that care needs to be employed when interpreting PHIP enhancements since the hydride signal strengths for two dihydride complexes with the same concentration need not be the same because they will reflect both their concentration and their ability to incorporate fresh parahydrogen. This means that the normalized signal intensities for 2 will only reflect the associated rate of introduction of fresh p-H2 into them. We therefore conclude that the extent of H2 exchange increases across the series PPh3 < P{p-tolyl}3 < P{C6H4-4-OMe}3 for 2a, 2b, and 2c. This trend parallels that associated with increase in electron donating ability of the phosphine.33 The anomalous behavior of the mixed phosphine systems reveals, however, that care needs to be taken before drawing any more precise conclusions about how the electronic properties of the phosphine relate to these observations. The relative enhancements of the hydride resonance of 2a relative to those of 3a and 3b, and 2a relative to those of 3a and 3c were also determined from these data. They proved to be 1:13.91:10.71 and 1:8:7.7, respectively. These data are not normalized since the 31P signals for 3 could not be seen, but they clearly confirm that the family of tris-pyridine (33) Tolman, C. A. Chem. ReV. 1977, 77, 313–348.

Table 2. Example Rate Constants for Equatorial Pyridine Ligand Exchange and Associated Activation Parameters for Exchange in Both 2 and 3 Rate complex (335 K)/s-1 2a 2b 2c 3a 3b

0.40 0.82 0.90 1.16 2.00

∆Hq/kJ mol-1

∆Sq/J mol-1 K-1

∆G295q/kJ mol-1

∆G335q/kJ mol-1

134 ( 1 125 ( 3 155 ( 9 91 ( 8 107 ( 3

151 ( 5 132 ( 9 222 ( 28 34 ( 25 85 ( 8

89.1 ( 0.2 86.2 ( 0.3 89.7 ( 1 81.1 ( 1.0 81.9 ( 0.2

83.1 ( 0.02 81.0 ( 0.03 80.8 ( 0.14 79.7 ( 0.05 78.5 ( 0.08

complexes 3 undergo significantly more rapid H2 exchange than that of 2. We have also examined analogous samples of 3a and 3b, and 3a and 3c at 330 K. The associated NMR spectra revealed that the relative signal enhancements for the hydride ligands of 3a/3b/3c were 2.25:1:0.79. It therefore appears that 3a undergoes more rapid exchange with H2 than either 3b or 3c. In this case therefore the extent of enhancement follows the reduction in electron donating power of the phosphine.33 We note that when EXSY spectroscopy was used to probe the hydride ligands in these complexes no exchange into free hydrogen was observed to occur in 2a on the NMR time scale at 335 K. In contrast, around 4% of the hydride signal intensity transferred into that of H2 when 3c was examined at 335 K. It is therefore clear that the reductive elimination of H2 from 2 or 3 under these conditions is not facile. Previous studies have suggested that H2 elimination proceeds dominantly from 16-electron complexes such as Ir(H)2Cl(PPh3)2 rather than 18-electron Ir(H)2Cl(PPh3)3.20 This process would therefore be expected to exhibit complex kinetic behavior, and given that its onset already pushes the temperature envelope of the solvent we have not explored it in more detail.20 Exchange of Bound and Free Pyridine Ligands by 2. EXSY spectroscopy was then employed to rigorously monitor the exchange rates of the bound pyridine ligand with free pyridine in 2 between 328 and 341 K. In all cases, the pyridine ligand that is trans to hydride was observed to undergo exchange with free pyridine. For 2a, the corresponding first order rate constant for pyridine dissociation proved to be 0.4 s-1 at 335 K, and values of ∆Hq ) 134 ( 1 kJ mol-1 and ∆Sq ) 151 ( 5 J mol-1 K-1 with ∆G295q ) 89.1 ( 0.2 kJ mol-1 were determined for this process. For 2b the corresponding rate constant proved to be 0.82 ( 0.008 s-1 at 335 K, and it can therefore be deduced that this process is faster in 2b than 2a. When this analysis method was repeated for 2c, the corresponding rate constant at 335 K proved to be 0.90 s-1. The associated activation parameters for this process are listed in Table 2. The contributions of both the ∆Sq and the ∆Hq terms are largest for the most electron rich phosphine, P{C6H4-4OMe}3 of 2c. However, their balance results in 2c having the greatest pyridine exchange rate at 335 K. In the case of the related complex, [Ir(NCMe)2(PPh3)2(H)2]+, the corresponding values are reported as 95.6 ( 2.3 kJ mol-1 and Inorganic Chemistry, Vol. 48, No. 2, 2009

667

Atkinson et al. -1

-1

28.2 ( 7.1 J mol K , respectively. The higher values of ∆Hq for pyridine loss in 2a, than acetonitrile in [Ir(NCMe)2(PPh3)2(H)2]+, agrees with the fact that the stronger pyridine donor should lead to the more costly rupture of a stronger Ir-N bond. 34

Exchange of Bound Pyridine and Free Pyridine in 3. A similar EXSY based procedure was used to study the exchange of pyridine in 3. This process proved to be faster than that found for 2. The associated rate constants and activation parameters for 3 are listed in Table 2. When 3a was examined using this approach the dominant pathway again corresponded to exchange of the equatorial pyridine ligands with free pyridine. At the higher temperatures, however, there was evidence for the intercoversion of the axial and equatorial pyridine ligands within 3a. At 340 K, and with a mixing time of 0.8 s, the proportion of pyridine loss was 60% while the level of axial-equatorial site exchange was only 4%. There were, however, no exchange peaks that connected the axial pyridine and free pyridine signals. These data therefore suggest that this process must occur via ligand interchange within the 16 electron intermediate [Ir(PPh3)(py)2(H)2]+ that is generated by pyridine loss trans to hydride in 3a rather than by direct dissociation of the pyridine ligand which is trans to phosphine. These observations also require that in the 16-electron intermediate [Ir(PPh3)(py)2(H)2]+, the two pyridine ligands remain inequivalent, a situation that would certainly hold true in the solvent complex [Ir(methanol)(PPh3)(py)2(H)2]+ which might be expected in methanol. No such complications were evident in the case of 3b, and the corresponding rates and activation parameters are listed in Table 2. In the case of 3b, the corresponding rate constant for pyridine loss is 2.00 s-1 at 335 K and therefore a factor of 2.4 faster than that of 2b. As a consequence the values of both ∆Hq and ∆Sq are reduced relative to those of 2. Much lower thermal stability was observed in the case of 3c with decomposition occurring over time at temperatures above 308 K. We measured a rate for pyridine dissociation at this temperature of 0.066 s-1 but were unable to obtain activation parameters because of the decomposition. This situation was further complicated because weak exchange into a new species which was tentatively assigned to [Ir(methanol)(P{C6H4-4-OMe}3) (py)2(H)2]+ was observed. When the Eyring data are used to estimate the corresponding rate constants at 308 K for 3a and 3b, the values obtained are 0.068 ( 0.016 s-1 and 0.064 ( 0.004 s-1, respectively. This confirms that at the lower temperatures, 3c undergoes comparable exchange to 3a, but clearly at the higher temperatures, where a large positive ∆Sq would work to enhance the difference, we cannot predict an order. There is a significant difference in the relative magnitudes of the activation entropies for 3 when compared to 2, and 3a when compared to 3b, even though they are essentially monitoring the same dissociation of pyridine. We note, however, that a higher level of associative character for the (34) Howarth, O. W.; McAteer, C. H.; Moore, P.; Morris, G. E. Dalton Trans. 1981, 1481–1485.

668 Inorganic Chemistry, Vol. 48, No. 2, 2009

Figure 3. (a) 1024 scan 15N control spectrum: (b) 16 scan 15N detected PHIP-INEPT spectrum during the reaction of 1b with parahydrogen in the presence of 15N-labeled pyridine in methanol-d4 (c) 64 scan 15N detected PH-INEPT+EXSY spectrum of the same sample

pyridine exchange step would lower ∆Sq, as would increased ordering of the phosphine in the TS. Oro et al. have determined that the activation parameters for acetonitrile dissociation from the site trans to hydride in fac-cis[Ir(NCMe)3(PiPr3)(H)2]+ are ∆Hq ) 112 ( 4 kJ mol-1 and ∆Sq ) 135 ) ( 8 J mol-1 K-1.27 The ∆Sq terms determined here for 3 are therefore not unusual. In the case of the monophosphine systems 3a and 3b, both the values of ∆Hq are consistently smaller than those found for their bis-phosphine analogues. This suggests the Ir-py bond energy in 2 is higher than that in 3 thereby indicating that the second phosphine acts to strengthen the Ir-N bond when compared to pyridine. We note that while both pyridine and phosphine are σ-donors and π-acceptors, phosphine is the higher field ligand. Magnetization Transfer from the PHIP Enhanced Hydride Signals to that of the 15N Centre in Bound Pyridine. A series of heteronuclear polarization transfer experiments were then tested for their ability to transfer the hydride polarizations exhibited by 2 and 3 into the 15N nuclei of the bound pyridine ligand. The NMR pulse sequences used were PH-INEPT, refocused PH-INEPT+, and PH-INEPT(+π/ 4) and have been reported previously.25 The associated NMR spectra were recorded for the same experimental setup and then the signal-to-noise ratios of the observed 15N resonances measured and normalized according to the relative signal strengths of the hydride resonances seen in the corresponding 1H NMR spectra that were recorded immediately before each 15N magnetization transfer experiment. Without the p-H2 based polarization transfer essentially no 15N signals were visible after 1024 scans as shown in Figure 3a. The most efficient transfer of magnetization proved to be achieved using the PH-INEPT pulse sequence, and the corresponding 16 scan experiment is illustrated in Figure 3b. This situation arises because in the refocused version of this experiment, PH-INEPT+, relaxation during the refocusing period results in only a third of the potential signal intensity surviving to the point of observation. The

Para-Hydrogen Induced Polarization Table 3. Comparison of 15N Signal to Noise Ratios for the Indicated Heteronuclear Polarization Transfer Experiments Employing Complex 1a experiment

signal/noisea

PH-INEPT 27.9 PH-INEPT+ 10.5 PH-INEPT(+π/4) 7.2 a Normalized using relative intensities of 1H hydride signals recorded immediately before the polarization transfer experiments were conducted.

Figure 5. Plot of magnetization, as a mole fraction, for the equatorial pyridine signal sites, starting from 3b versus reaction time obtained from PH-INEPT+EXSY data for a sample prepared from 1b, 15N-pyridine and p-H2 at 330 K.

Figure 4. Schematic of the pulse sequence PH-INEPT+EXSY used for magnetization transfer where (τ ) 1/4JHN).

PH-INEPT(+π/4) experiment was found to be least effective. These data are summarized in Table 3. Magnetization Transfer from the PHIP Enhanced Hydride Signals to that of the 15N Centre in Free Pyridine. Our main goal in these experiments was to transfer p-H2 derived polarization from bound to free pyridine in a second step. This process would therefore lead to the polarization of a substrate that did not contain any protons that originated in the source of polarization, p-H2. We therefore modified these polarization transfer experiments to store the 15N magnetization in an Iz term prior to the introduction of a variable delay to facilitate magnetization transfer into the signal for free pyridine. This situation can be simply thought of as a heteronuclear EXSY experiment. The new pulse sequence, PH-INEPT+EXSY, is illustrated in Figure 4 where the polarization transfer sequence is followed by a 90°-y storage pulse, and the magnetization transfer delay (d) is followed by a 90° observation pulse. When this process was carried out on a sample containing enhanced proton signals for 2a and 3a it became clear that while polarization transfer to bound 15N can be achieved, little or no exchange of magnetization corresponding to the ligated ligand exchanging into free pyridine occurs at 330 K. This is in direct agreement with the quantitative results described earlier. The PH-INEPT+EXSY based data collection was then repeated on samples containing both 1b and 1c. The 15N spectra now show significant signals for both bound pyridine and free pyridine. As the delay time used to encode ligand dissociation is increased from 0.1 to 1 s the proportion of signal that resides at the position of free pyridine clearly increases at the expense of the signals for the bound pyridine ligands of 3b. These data are illustrated as a time course plot in Figure 5. A 64 scan spectrum with a reaction delay of 1 s is illustrated in Figure 3c for a system based on the precursor 1b. The signal-to-noise ratio for the pyridine resonance in this 64 scan experiment is 198.4 and compares with the 1024 scan control experiments value of 6.64. This equates to a 120-fold increase in signal strength via polarization transfer.

In the corresponding control experiments, using INEPT, and PH-INEPT sequences, no signal for free pyridine can be observed on this time scale. The control experiments therefore confirm that it is magnetization transfer from p-H2 derived hyperpolarization sitting on the hydride ligands of 3b that is responsible for these observations rather than magnetization transfer from the ring protons of free pyridine. These results represent the successful demonstration of the PHIP sensitization of an organic molecule that contains no nuclei that were previously located in parahydrogen. Summary In this paper we have demonstrated that [Ir(PR3)2(py)2(H)2]BF4 and fac, cis-[Ir(PR3)(py)3(H)2]BF4 where R ) Ph, p-tolyl and p-methoxyphenyl, undergo pyridine loss in a process that is influenced by the identity of the phosphine. The rate of pyridine ligand dissociation at 335 K proved to be highest for the most electron rich p-methoxyphenyl phosphine in the bis-pyridine complexes 2 which corresponds to the situation where both ∆Hq and ∆Sq are maximized. In contrast for 3, the rate constants for pyridine loss in the tris-pyridine systems are all similar at 308 K, although they are larger than those determined for 2 (Table 2). When these complexes are prepared using p-H2 and 15Nlabeled pyridine was employed as the substrate, the parent complexes 2 and 3 show substantial PHIP enhancements in their hydride resonances. Magnetization transfer from the enhanced hydride signals of these species has been demonstrated to provide access to substantial 15N signal enhancements in the bound pyridine ligand via the application of an INEPT based protocol. When such an approach is used in conjunction with a magnetization transfer step, the hyperpolarization that was originally based in p-H2 is transferred first to a bound pyridine ligand in fac, cis-[Ir(PR3)(py)3(H)2]BF4 and then to free pyridine itself. This relayed magnetization transfer demonstrates that it is possible to employ the p-H2 to sensitize the NMR signals of materials that have not been functionalized by an H2 addition reaction. We have estimated that a 120-fold gain in the signal strength of the free pyridine resonances can be achieved via this route. It should be noted that since this means 1 scan equates to 14,400 scans without polarization there is therefore a significant reduction in the time necessary to record such information. Furthermore, since sensitivity relates to the square of the size of the magnetic Inorganic Chemistry, Vol. 48, No. 2, 2009

669

Atkinson et al. field, similar results could be achieved on a 103 T magnet which is around five times larger than that found in a 900 MHz system. Previously, reactions with p-H2 have been shown to enable the detection of metal dihydride complexes. It is not surprising, however, that when the protons that were previously located in p-H2 are transferred into a hydrogen-acceptor such as an alkene, strong observable magnetization is also produced in the corresponding hydrogenation products. Recently this approach has been used to prepare hyperpolarized organic molecules for use as contrast agents in magnetic resonance imaging (MRI).35 However, the need for the existence of a suitable hydrogen-acceptor reflects a significant limitation in the existing p-H2 approach. The novelty of the procedure illustrated here is that it does not require the functionalization of a substrate through a formal hydrogenation step. Consequently it offers the opportunity to sensitize substrates where there is no dehydro-analogue available. We are continuing to explore the potential of this and other routes to sensitize the NMR experiment more generally. Experimental Details NMR measurements were made on Bruker DRX400 (1H at 400.13 MHz, 31P at 161 MHz, 15N at 40 MHz), Bruker AV500 (1H at 500.22 MHz, 31P at 202.50 MHz, 15N at 50.69 MHz), Bruker wide-bore AV600 (1H at 600.12 MHz, 31P at 242.93 MHz, 15N at 60.81 MHz) and Bruker AV 700 (1H at 700.12 MHz) NMR spectrometers. NMR samples were made up in 5 mm Young’s tapped NMR tubes. Typical samples contained 1 mg of the iridium complex dissolved in CD3OD (600 µL) with the other reagents being added by microsyringe. Samples were degassed at -78 °C and then pressurized with 3.5 bar p-H2 before appropriate NMR measurements were recorded. p-H2 was prepared by cooling the gas to 20 K over activated charcoal as described previously.20 X-ray data were obtained for a crystal of 2a at 110 K using a Bruker Smart Apex diffractometer with Mo KR radiation (ν ) 0.71073 Å) in conjunction with a SMART CCD camera. Diffractometer control, data collection, and initial unit cell determination were performed using “SMART” (v5.625 Bruker-AXS). Frame integration and unit-cell refinement were carried out with the “SAINT+” (v6.22, Bruker AXS) package. Absorption corrections were applied by SADABS (v2.03, Sheldrick). The structure was solved by Patterson map using ShelXS (Sheldrick, 1997) and refined using ShelXL (Sheldrick, 1997). Crystals were grown from a benzene solution of 1a which contained H2 and pyridine. Syntheses. All reactions were carried out under N2 using glovebox or Schlenk line techniques. Dichloromethane and diethyl ether were distilled over calcium hydride and sodium benzophenyl ketyl, respectively, prior to use. Tri(phenyl)phosphine (Aldrich), tri(p-tolyl)phosphine (Strem), tri(anisolyl)phosphine (Strem), pyridine (Aldrich), and 15N-pyridine (Cambridge Isotopes) were used (35) Goldman, M.; Johannesson, H.; Axelsson, O.; Karlsson, M. C. R. Chim. 2006, 9, 357–363.

670 Inorganic Chemistry, Vol. 48, No. 2, 2009

as supplied and stored under N2. [Ir(COD)(PPh3)2]BF4 (1a) and [Ir(COD)(NCMe)2]BF4 were prepared by literature methods.26,29,36,37 [Ir(COD)(P{p-tolyl}3)2]BF4 (1b). The known complex 1b was prepared as follows: [Ir(COD)(NCMe)2]BF4 (150 mg, 0.32 mmol) was dissolved in dichloromethane (5 mL) to which was added by canula P(p-tolyl)3 (195 mg, 0.64 mmol) in dichloromethane (5 mL). After stirring for 15 min, the volume was reduced in vacuo and diethyl ether added to precipitate the product. The dark pink/red solid was then washed with diethyl ether (5 mL) and dried under vacuum. Yield: 197 mg (62%): 31P NMR: δ 16.1. [Ir(COD)(P{C6H4-4-OMe}3)2]BF4 (1c). The known complex 1c was prepared as follows: A Schlenk tube was charged with [Ir(COD)Cl]2 (250 mg, 0.37 mol) and AgBF4 (160 mg, 0.82 mmol). To this was added a degassed methanol (10 mL) solution of P(C6H44-OMe)3 (521 mg, 1.48 mmol), and the resulting suspension stirred overnight in the absence of light. The solvent was then completely removed under vacuum, and the product extracted into CH2Cl2 (10 mL). The suspension was filtered through celite, and the red filtrate concentrated to about 1 mL before Et2O (10 mL) was added to precipitate the product. After washing with Et2O (3 × 2 mL) and drying in vacuo for 1 h a dark red powder was obtained. Yield: 640 mg, 79%. 31P NMR: δ 14.9 Exchange of IrH for IrD. It should be noted that during the course of all of these reactions in CD3OD, the slow exchange of a deuterium label into the hydride sites of 2 and 3 was observed. After 24 h, the corresponding hydride resonances were very weak, or no longer visible, because of total exchange to the corresponding deuterides. Both CD3OH and HD signals were also visible in these spectra. Selected NMR data for [Ir(py)2(PPh3)(P{p-tolyl}3)(H)2]BF4. 1 H: Hydride: δ -21.72, (JPH ) 19 Hz, JNH ) 23 Hz); 31P: PPh3 δ 23.17 (d JPP ) 344 Hz), P{p-tolyl}3, δ 20.47 (d, JPP ) 344 Hz), 15 N: δ -63.3. Selected NMR data for [Ir(py)2(PPh3)(P{C6H4-4-OMe}3)(H)2] BF4. 1H: Hydride: δ -21.72, (JPH ) 21 Hz, JNH ) 21 Hz); 31P: PPh3 δ 22.86 (d JPP ) 348 Hz), P{C6H4-4-OMe}3, δ 18.35 (d, JPP ) 348 Hz), 15N: δ -61.1.

Acknowledgment. We wish to thank the University of York and a White Rose initiative for part funding this work through its technology transfer scheme. Other financial support from the Spanish MEC (Project Consolider ORFEO (CSD 2007-00006)) and the MRC (K.D.A.) are also acknowledged. We are also grateful to Bruker BioSpin and Mark Mortimer for particularly helpful discussions. Supporting Information Available: Tables containing rates and activation parameters for pyridine exchange and X-ray information for 2a (CCDC 699760). This material is available free of charge via the Internet at http://pubs.acs.org. IC8020029 (36) Crabtree, R. H.; Mellea, M. F.; Mihelcic, J. M.; Sen, A.; Chebolu, V. Inorg. Synth. 1989, 26, 122–126. (37) Crabtree, R. H.; Hlatky, G. G.; Parnell, C. P.; Segmuller, B. E.; Uriarte, R. J. Inorg. Chem. 1984, 23, 354–358.